CN107612016B - Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy - Google Patents
Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy Download PDFInfo
- Publication number
- CN107612016B CN107612016B CN201710671973.1A CN201710671973A CN107612016B CN 107612016 B CN107612016 B CN 107612016B CN 201710671973 A CN201710671973 A CN 201710671973A CN 107612016 B CN107612016 B CN 107612016B
- Authority
- CN
- China
- Prior art keywords
- distributed power
- distribution network
- power supply
- voltage
- population
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a planning method of a distributed power supply in a power distribution network based on maximum voltage correlation entropy, which comprises the steps of establishing a planning model of the distributed power supply with the best economic index and the best improved voltage quality as optimization targets, and researching the optimal access position and the optimal access capacity of the distributed power supply by selecting a Memetic algorithm. According to the planning method, the maximum correlation entropy is introduced as a cost function of the voltage quality, the influence of noise on the node voltage can be weakened, the node voltage of the system is well reflected, and therefore the planning of the distributed power supply is facilitated. The Memetic algorithm is used as the optimization algorithm, the optimization result is better than that obtained by the traditional optimization algorithm, and the algorithm efficiency is higher.
Description
Technical Field
The invention belongs to the technical field of distributed power supply planning methods, and particularly relates to a planning method for a distributed power supply in a power distribution network based on a maximum voltage correlation entropy.
Background
With the continuous increase of global energy demand, the continuous exhaustion of resources and the continuous deterioration of climate, the three major problems are highlighted, so that renewable energy plays a very important role in energy strategy of each country. Among them, distributed power sources (such as wind power generation, photovoltaic power generation, etc.) have been rapidly developed due to their cleanliness, environmental friendliness, and low investment. Distributed Generation (DG) refers to an efficient, reliable, clean power generation unit with small capacity and arranged in the vicinity of the user. When a large number of distributed power supplies are connected to a power distribution network, node voltage, network loss, reliability and the like are obviously affected, and the influence degree is closely related to the factors such as the positions and capacities of the distributed power supplies connected to the power distribution network. Therefore, the access of the distributed power supply in the power distribution network is planned and optimized, so that natural resources can be fully utilized, and the method has very important basic effects on the aspects of improving the safety and the economy of the power distribution network.
Planning optimization of the distributed wind power supply is deeply researched at home and abroad. The planning model mainly aims at the minimum of fixed investment and annual operation cost of the wind power plant, the influence of the access of the distributed power supply on the safety of the power grid cannot be considered, the calculation tends to be conservative, and the access capacity of the distributed wind power supply is limited. Some documents aim at the maximum access capacity of the distributed power supply, but do not consider the influence of unstable voltage, large fluctuation and the like caused by excessive access. For the voltage quality problem, the voltage deviation is mostly used as an objective function, and the cost function is all expressed by Mean Square Error (MSE), but in an actual system, the voltage value change is caused by some indeterminate factors. The solution of the distributed wind power supply planning mostly adopts an intelligent optimization algorithm, such as a genetic algorithm, a particle swarm algorithm and the like, and the algorithms inevitably fall into local optimization easily, are low in solution speed, and are not converged in iteration.
Disclosure of Invention
The invention aims to provide a method for planning distributed power supplies in a power distribution network based on the maximum voltage correlation entropy, which breaks the limitation of mean square error in processing non-Gaussian nonlinear noise on the basis of improving the voltage quality and reducing the economic index, and solves the problem that the planned distributed power supplies in the conventional power distribution network cannot be considered and coordinated.
The technical scheme adopted by the invention is that the method for planning the distributed power supply in the power distribution network based on the maximum voltage correlation entropy comprises the following steps:
(a) taking the operation cost and the voltage quality of the power distribution network as an optimization objective function;
(b) a function of the operating cost of the distribution network;
(c) a function of voltage quality;
(d) determining constraint conditions including power flow constraint of a power distribution network, equipment capacity constraint, node voltage constraint and limitation of a distributed power supply;
step 2.1, Generation of initial population
Before solving, representing feasible solution data of a solution space such as a result of distributed power supply address selection and constant volume into array structure data of a search space such as an access node number and distributed power supply capacity corresponding to the node number, wherein different feasible solutions are formed by different combinations of the array structure data, and the number of initial populations is set to be M;
selecting function dimensionality as at least two dimensions, wherein one dimension represents an access node, and the other dimension represents access capacity;
using each feasible solution as an individual of the population, and applying a limited initial population generation method to directly limit the output of a power supply within the maximum range of the total capacity of the selected test system in the population generation process to obtain an initial population;
step 2.2, interleaving
Randomly selecting two individuals in the initial population generated in the step 2.1, and performing operation through a crossover operator to generate two new individuals of a new generation population to form a new population;
step 2.3, mutation
Selecting any plurality of individuals from the new population generated after the operation of the step 2.2, and carrying out mutation operation according to a mutation operator to obtain a mutation population;
step 2.4, fitness calculation
Substituting the variant population in the step 2.3, namely the node number, into the planning model in the step 1, and obtaining the operation cost and the voltage quality of the corresponding power distribution network, namely the maximum correlation entropy of the node voltage through load flow calculation to form a population to be selected;
step 2.5, selection
Discarding the individuals with low fitness in the step 2.4, selecting M individuals with optimal corresponding optimization objective functions from the group to be selected in the step 2.4, and selecting the individuals to enter the next iteration process according to the selection probability;
step 2.6, local search
Performing local search on all individuals selected in the step 2.5 by adopting a pure method;
step 2.7, checking whether the optimizing result meets the optimizing condition through the local search in the step 2.6:
if the optimal solution is met, stopping, namely obtaining the optimal solution;
if not, repeating the steps 2.2-2.6 until the optimal scheme is obtained.
The present invention is also characterized in that,
the optimization objective function in step 1(a) is:
c in formula (1)allFor the running cost of the distribution network, UMCCIs the voltage quality, i.e. the maximum associated entropy of the node voltage.
The function of the operating cost of the power distribution network in step 1(b) is:
Call＝CL+CDG+Cpur(2)
CL＝Ce·Ploss·TLmax(3)
Cpur＝Ce(PLA-P∑DG+Ploss)TLmax(5)
wherein C in the above formulaeallFor the operating costs of the distribution network, CLFor loss of distribution network, CDGFor the total cost of operation of the distributed power supply, CpurTo total cost of electricity purchase, CeRepresents unit electricity price, yuan/kW.h, PlossRepresents the loss of the network, TLmaxRepresents the number of annual hours of maximum load, PGenIs rated active power of the distributed power supply, CeDGCost per unit of electricity for distributed power supply, TDGmaxIs the maximum number of hours of power generation, M, of the distributed power supplyDGTotal number of distributed power sources, P, for access to the distribution networkLATo total capacity of the grid, P∑DGIs the total active output of the distributed power supply.
The function of the voltage quality in step 1(c) is:
wherein, UMCCIs the voltage quality, i.e. the maximum associated entropy of the node voltage, n is the number of nodes, UiSigma is a kernel function parameter, which is a per unit value of each node voltage.
The function of the power flow constraint of the power distribution network in the step 1(d) is as follows:
wherein P isiFor active injection at node i, QiFor reactive injection at node j, UiAnd UjThe voltage amplitudes at nodes i, j, G, respectivelyijConductance for branch ij, BijSusceptance, θ, of branch ijijIs the voltage phase angle difference between nodes i and j;
the function of the equipment capacity constraint in step 1(d) is:
wherein P isDGiFor active power of distributed power, QDGiIs the reactive power output of the distributed power supply,is the rated active power of the distributed power supply,rated reactive power for the distributed power supply;
the function of the node voltage constraint in step 1(d) is:
Uimin≤Ui≤Uimax(9)
wherein U isiIs the node voltage, UiminIs the lower limit of the node voltage, UimaxIs the upper node voltage limit.
The crossover operator in step 2.2 is as follows:
wherein P is1And P2Two father individuals randomly selected from the initial population respectively;
the filial generation generated after the operation of the cross operator corresponds to the new individual; omega1、ω2Is [0, 1]]And (4) selecting the parameters randomly.
The mutation operator in step 2.3 is specifically:
wherein V is selected variant individual, V' is variant individual, sign randomly takes 0 or 1, bsupAnd binfRespectively an upper bound and a lower bound of the parameter value, r is [0, 1]]A random number generated above;
for population evolution markers, where gcIs the current evolutionary algebra of the population, gmIs the maximum evolutionary algebra of the population.
Selection probability P in step 2.5iThe following formula:
wherein JiAnd the fitness value corresponding to the population i.
The planning method of the invention has the beneficial effects that:
1) compared with the existing voltage mean square error model, the model in the planning method better reflects the voltage quality of the system, and can better promote the node voltage of the configured system in planning;
2) compared with the traditional particle swarm optimization, the optimization of the model in the planning method is quicker and more efficient, and the optimal configuration scheme of the multipoint access of the distributed power supply is easier to realize.
Drawings
FIG. 1 is a flow chart of distributed power planning using Memetic algorithm of the present invention;
fig. 2 is an example of IEEE33 node calculation in embodiment 1 of the present invention;
FIG. 3 is a comparison graph of voltages before and after different numbers of distributed power supplies are connected in the present invention;
FIG. 4 is a voltage comparison graph of the MCC and MSE planning models of the present invention;
FIG. 5 is a voltage comparison graph of the particle swarm planning model and the memetic planning model in the present invention;
FIG. 6 is an iterative process of Memetic algorithm in the present invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention relates to a method for planning a distributed power supply in a power distribution network based on maximum voltage correlation entropy, which comprises the following steps:
(a) and taking the running cost and the voltage quality of the power distribution network as an optimization objective function. In order to ensure the safe and reliable operation of the power grid, the power flow constraint, the equipment capacity constraint and the node voltage constraint must be met, the planning method of the invention adopts the maximum correlation entropy of the distribution network operation cost and the voltage deviation of the operation cost and the network loss cost of the added distributed power supply as a target function from the perspective of a power supplier, establishes a multi-target optimization model with constraint conditions, such as a formula (1),
wherein C isallFor the running cost of the distribution network, UMCCThe voltage quality, namely the maximum correlation entropy of the node voltage;
(b) a function of the operating cost of the distribution network, such as equations (2) - (5),
Call＝CL+CDG+Cpur(2)
CL＝Ce·Ploss·TLmax(3)
Cpur＝Ce(PLA-P∑DG+Ploss)TLmax(5)
wherein C in the above formulaeallFor the operating costs of the distribution network, CLFor loss of distribution network, CDGFor the total cost of operation of the distributed power supply, CpurTo total cost of electricity purchase, CeRepresents unit electricity price, yuan/kW.h, PlossRepresents the loss of the network, TLmaxRepresents the number of annual hours of maximum load, PGenIs rated active power of the distributed power supply, CeDGCost per unit of electricity for distributed power supply, TDGmaxIs the maximum number of hours of power generation, M, of the distributed power supplyDGTotal number of distributed power sources, P, for access to the distribution networkLATo total capacity of the grid, P∑DGIs the total active output of the distributed power supply.
(c) The function of the voltage quality is as in equation (6),
wherein, UMCCIs the voltage quality, i.e. the maximum associated entropy of the node voltage, n is the number of nodes, UiSigma is a kernel function parameter, which is a per unit value of each node voltage.
In the maximum correlation entropy criterion, a kernel function k meeting the Mercer criterion needs to be introducedσThe kernel function is here chosen to be σ2σ is a kernel function parameter;
where T is the data length, d is the data dimension, σxIs the standard deviation of the data; in practice, according to different situations, different kernel function parameters are selected, and better performance can be obtained.
The voltage quality expression in the distributed power supply planning model needs to process non-Gaussian nonlinear noise; in order to well process the problem of error distribution with non-Gaussian characteristics, a new similarity evaluation function-generalized correlation entropy is provided by combining the concept of entropy and a correlation function in information theory in the neural network laboratory of the university of Florida; according to the concept, a Maximum correlation entropy Criterion (MCC) is induced, and the MCC is widely applied to the fields of adaptive filtering, mode classification, dimension reduction operation, feature selection and the like when used as a cost function.
In other distributed power supply planning designs, the root mean square error (U) of the node voltage is often usedMSE) As a reference standard for measuring the voltage quality, the maximum correlation entropy (U) of the node voltage is used in the planning method of the inventionMCC) To replace the common root mean square error (U) of the node voltageMSE) As a cost function, U can be brokenMSELimitations in dealing with non-gaussian non-linear noise. Unlike the global similarity measure criterion MSE, MCC focuses more on local similarity because the value of the correlation entropy is mainly determined by a kernel function that is positively correlated with the straight line a ═ B; therefore, the MCC is adopted to replace MSE as a target function, so that the local deviation of voltage distribution can be effectively reflected, the influence of a mutation point on an optimization result in the algorithm operation process is weakened, and the overall situation can be guaranteed at the same time of site selection and volume fixing. In addition, when the voltage fluctuation is large, the voltage can better follow the fluctuation point, and the planning scheme is ensured to have better electric energy quality.
(d) Determining constraint conditions including power flow constraint of a power distribution network, equipment capacity constraint, node voltage constraint and limitation of a distributed power supply;
step 1(d) the function of the power flow constraint of the power distribution network is as follows:
wherein P isiFor active injection at node i, QiFor reactive injection at node j, UiAnd UjThe voltage amplitudes at nodes i, j, G, respectivelyijConductance for branch ij, BijSusceptance, θ, of branch ijijIs the voltage phase angle difference between nodes i and j;
the function of the equipment capacity constraint in step 1(d) is:
wherein P isDGiFor active power of distributed power, QDGiIs the reactive power output of the distributed power supply,
is the rated active power of the distributed power supply,
rated reactive power for the distributed power supply;
the function of the node voltage constraint in step 1(d) is:
Uimin≤Ui≤Uimax(9)
wherein U isiIs the node voltage, UiminIs the lower limit of the node voltage, UimaxIs the upper node voltage limit.
step 2.1, Generation of initial population
Before solving, firstly, representing feasible solution data of a solution space such as a result of distributed power supply address selection and constant volume into array structure data of a search space such as an access node number and distributed power supply capacity corresponding to the node number, wherein different feasible solutions are formed by different combinations of the array structure data, and adding an initial population set as M;
selecting function dimensionality to be at least two-dimensional, wherein one dimension represents an access node, the other dimension represents access capacity, and the dimensionality of the function can be expanded if multi-point access to the distributed power supply is to be realized;
and (3) taking each feasible solution as an individual of the population, and applying a limited initial population generation method to directly limit the output of a power supply within the maximum range of the total capacity of the selected test system in the population generation process to obtain an initial population.
Step 2.2, interleaving
And 2.1, randomly selecting two individuals in the initial population generated in the step 2.1, and performing operation through a crossover operator to generate two new individuals of a new generation population to form a new population, wherein the crossover operator has the following formula:
wherein P is1And P2Two father individuals randomly selected from the initial population respectively;the filial generation generated after the operation of the cross operator corresponds to the new individual; omega1、ω2Is [0, 1]]And (4) selecting the parameters randomly.
Step 2.3, mutation
Selecting any several individuals from the new population generated after the operation of the step 2.2 according to a certain mutation probability, and carrying out mutation operation according to a mutation operator to obtain a mutation population; the mutation operator is specifically:
wherein V is selected variant individual, V' is variant individual, sign randomly takes 0 or 1, bsupAnd binfRespectively an upper bound and a lower bound of the parameter value, r is [0, 1]]A random number generated above;
for population evolution markers, where gcIs the current evolutionary algebra of the population, gmIs the maximum evolutionary algebra of the population.
Step 2.4, fitness calculation
And (3) substituting the variation population in the step (2.3), namely the number of the node and the capacity of the distributed power supply corresponding to the node into the planning model in the step (1), and obtaining the operation cost and the voltage quality of the corresponding power distribution network, namely the maximum correlation entropy of the node voltage through load flow calculation to form a population to be selected.
Step 2.5, selection
Abandoning the individuals with low fitness in the step 2.4, and selecting M individuals with optimal corresponding optimization objective functions from the population to be selected in the step 2.4 so as to lead the individuals to be selected according to the selection probability PiIs selected to enter the next iteration process, and the probability P is selectediThe following formula:
wherein JiAnd the fitness value corresponding to the population i.
Step 2.6, local search
Performing local search on all individuals selected in the step 2.5 by adopting a pure method;
step 2.7, checking whether the optimizing result meets the optimizing result through the local search in the step 2.6:
if the optimal solution is met, stopping, namely obtaining the optimal solution;
if not, repeating the steps 2.2-2.6 until the optimal scheme is obtained.
Examples
Establishing a planning model according to the step 1, and solving the planning model by using a Memetic algorithm to obtain an optimal scheme, specifically as shown in FIG. 1, including the following steps:
step 1) Generation of initial population
The system voltage in example 1 was 12.66kV, the total load of the system was 5084.26+ j2547.32kva, and the power network loss was 35.36 kW. The total load of the system of the example is 5.084MW, and the proportion of power supply access does not exceed the total load of the system, so the generation of the initial population of the algorithm is random numbers on [0, 5MW ], as shown in FIG. 2, and the node constraint condition is 1 ~ 33. The function dimension is selected to be two-dimensional, one dimension represents an access node, and the other dimension represents access capacity (if the distributed power supply is accessed in multiple points, the function dimension can be expanded). The population number 20 is set, and the number of iterations is 20.
Step 2) Cross strategy
Setting the crossing probability to be 0.95, and calculating new individuals after the initial population crossing according to a formula (10) pair.
Step 3) mutation strategy
Setting the mutation probability to be 0.1, randomly generating a random number r on [0, 1], and if r is smaller than the mutation probability, mutating the crossed individuals according to a formula (11).
Step 4) fitness function
Randomly generating M groups of chromosomes (namely randomly generating M groups of node numbers and node capacities corresponding to the node numbers)]) M different network topologies are formed respectively, and a Newton method is adopted to perform load flow calculation on the M new network topologies. After load flow calculation is carried out on M new network topologies, objective function values such as power grid operation cost, maximum voltage correlation entropy and the like under each network topology are obtained according to a formula (2) and a formula (6), normalization processing is carried out on the objective function values, weight parameters are set, and calculation results are taken into the formula F-k1C'all-k2U'MCCCalculating to obtain a fitness value F;
wherein C 'in the formula'allIs the operation cost of the power distribution network after the normalization at this time, U'MCCThe voltage quality after the normalization, namely the maximum correlation entropy, k, of the node voltage1Is the running cost C of the power distribution network'allWeight coefficient of (1), k2Node voltage maximum correlation entropy U'MCCThe weight coefficient of (2).
Step 5) selecting a policy
And (4) calculating the selection probability of each individual according to the formula (12), if the individual selection probability is greater than a random number, storing the individual, and otherwise, removing the individual.
Step 6) local search strategy
And local search is carried out on all individuals in the population by adopting a pure method.
Step 7) judging whether the condition of optimizing convergence is met
Checking whether the difference value of two adjacent optimization results is 10-5Within the range, the iteration times are more than 20, and if yes, the result is output; if not, continuously and repeatedly turning to the steps 2) -6) to continuously execute until the convergence condition is met.
Refer to the average on-line electricity price of China, CeTaking 0.3 yuan/(kWh), the annual maximum load utilization hours TLmaxIs 3000h, CeDGTaking 0.354 yuan/(kWh), TDGmaxTake 2400h, k1For distribution network operating costs CallHas a weight coefficient of 0.5, k2For maximum associated entropy U of voltage deviationMCCThe weight coefficient of (2) is 0.5. The value of the kernel function parameter sigma is calculated by using the Silverman criterion, but the calculated sigma is too small, and the selection of the kernel function parameter sigma should be 10-3～103So the kernel parameter σ is 0.001.
The following simulation calculation of the IEEE33 node example is performed by using the memetic algorithm according to the established model of the present application. Table 1 shows the configuration calculation results.
TABLE 1 optimization results
The network loss of the distribution network without the distributed wind power supply is 0.3496MW, and it can be seen from table 1 that the network loss of the 33-node network accessing two distributed power supplies is reduced by 72.71% compared with the network loss without the distributed power supplies, and the network loss accessing four distributed power supplies is reduced by 84.81% compared with the network loss without the distributed power supplies, which indicates that the network loss can be effectively reduced by reasonably accessing the distributed wind power supply. As can be seen from the node voltage comparison graph in fig. 3, the node voltage of the system after the distributed power supplies are connected is obviously improved, but the network loss and the node voltage after the four DGs are connected are ideal compared with the network loss and the node voltage after the two distributed power supplies are connected, so that the network loss and the voltage quality in the distribution network can be reduced and improved by properly connecting more distributed wind power supplies, but the cost is increased, and therefore, investors can make decisions by considering various factors during planning.
In the planning method in the application, the maximum correlation entropy (U) of the node voltage is usedMCC) To replace the root mean square error (U) of the node voltage in generalMSE) As a cost function, break UMSELimitations in dealing with non-gaussian non-linear noise.
Next, to verify the U selected in this applicationMCCThe superiority of the objective function is expressed as UMCCAnd UMSEThe distributed wind power supply is planned and designed for the objective function, and the comparative analysis result is shown in table 2:
TABLE 2 planning results
As can be seen from the optimization results in Table 2 and FIG. 4, U is usedMCCThe method is used for planning and designing as a cost function for describing the voltage deviation, the network loss of the obtained configuration scheme is lower, and meanwhile, the method has obvious advantages in the aspect of improving the voltage quality. When the distributed wind power supply of the power distribution network is planned and designed, the voltage maximum correlated entropy value is used for replacing the voltage root mean square error to serve as a target function, the influence of a mutation point on an optimization result in the algorithm operation process can be weakened, and the situation that the whole situation can be considered when site selection and volume fixing are carried out is guaranteed. In addition, when the voltage fluctuation is large, the maximum voltage deviation correlation entropy is taken as a target function, so that fluctuation points can be better tracked, and a planning scheme is ensured to have better electric energy quality.
Also for the IEEE33 node system, a particle swarm function is used for solving a planning model, and compared with a configuration result of a culture evolution algorithm, the analysis is as follows:
TABLE 3 optimization results
The Memetic algorithm adopts a calculation frame and an operation flow similar to a genetic algorithm, but is not limited to a simple genetic algorithm, local search is carried out on the algorithm after each crossing and variation, and the convergence capability of the algorithm is accelerated by eliminating bad individuals as soon as possible to optimize a population structure. As is clear from table 3 and fig. 5, the programming result obtained by the memetic algorithm is smaller in both the grid loss and the fitness value than the programming value obtained by the particle swarm algorithm, and the obtained voltage result is also better. In addition, the particle swarm algorithm needs to iterate for 100 times to obtain the optimization result, and as shown in fig. 6, the Memetic algorithm can quickly obtain the optimization result, and the optimization result tends to be stable within 20 times on average, so the Memetic algorithm has more advantages than the traditional particle swarm algorithm.
Claims (8)
1. The method for planning the distributed power supply in the power distribution network based on the maximum voltage correlation entropy is characterized by comprising the following steps of:
step 1, establishing a planning model of a distributed power supply in a power distribution network, wherein the planning model comprises the following steps:
(a) taking the operation cost and the voltage quality of the power distribution network as an optimization objective function;
(b) a function of the operating cost of the distribution network;
(c) a function of voltage quality;
(d) determining constraint conditions including power flow constraint of a power distribution network, equipment capacity constraint, node voltage constraint and limitation of a distributed power supply;
step 2, analyzing and processing the planning model established in the step 1 by using a Memetic algorithm so as to determine the position and the capacity of the distributed power supply accessed to the power distribution network and obtain an optimal scheme, and specifically comprising the following steps:
step 2.1, Generation of initial population
Before solving, representing feasible solution data of a solution space such as a result of distributed power supply address selection and constant volume into array structure data of a search space such as an access node number and distributed power supply capacity corresponding to the node number, wherein different feasible solutions are formed by different combinations of the array structure data, and the number of initial populations is set to be M;
selecting function dimensionality as at least two dimensions, wherein one dimension represents an access node, and the other dimension represents access capacity;
using each feasible solution as an individual of the population, and applying a limited initial population generation method to directly limit the output of a power supply within the maximum range of the total capacity of the selected test system in the population generation process to obtain an initial population;
step 2.2, interleaving
Randomly selecting two individuals in the initial population generated in the step 2.1, and performing operation through a crossover operator to generate two new individuals of a new generation population to form a new population;
step 2.3, mutation
Selecting any plurality of individuals from the new population generated after the operation of the step 2.2, and carrying out mutation operation according to a mutation operator to obtain a mutation population;
step 2.4, fitness calculation
Substituting the variant population in the step 2.3, namely the node number, into the planning model in the step 1, and obtaining the operation cost and the voltage quality of the corresponding power distribution network, namely the maximum correlation entropy of the node voltage through load flow calculation to form a population to be selected;
step 2.5, selection
Discarding the individuals with low fitness in the step 2.4, selecting M individuals with optimal corresponding optimization objective functions from the group to be selected in the step 2.4, and selecting the individuals to enter the next iteration process according to the selection probability;
step 2.6, local search
Performing local search on all individuals selected in the step 2.5 by adopting a pure method;
step 2.7, checking whether the optimizing result meets the optimizing convergence condition through the local search in the step 2.6:
if the optimal solution is met, stopping, namely obtaining the optimal solution;
if not, repeating the steps 2.2-2.6 until the optimal scheme is obtained.
2. The method for planning a distributed power source in a power distribution network based on maximum voltage correlation entropy of claim 1, wherein the optimization objective function in the step 1(a) is as follows:
c in formula (1)allFor the running cost of the distribution network, UMCCIs the voltage quality, i.e. the maximum associated entropy of the node voltage.
3. The method for planning a distributed power supply in a power distribution network based on maximum voltage correlation entropy of claim 1, wherein the function of the operation cost of the power distribution network in the step 1(b) is as follows:
Call＝CL+CDG+Cpur(2)
CL＝Ce·Ploss·TLmax(3)
Cpur＝Ce(PLA-PΣDG+Ploss)TLmax(5)
wherein C in the above formulaeallFor the operating costs of the distribution network, CLFor loss of distribution network, CDGFor the total cost of operation of the distributed power supply, CpurTo total cost of electricity purchase, CeRepresents unit electricity price, yuan/kW.h, PlossRepresents the loss of the network, TLmaxRepresents the number of annual hours of maximum load, PGenIs rated active power of the distributed power supply, CeDGCost per unit of electricity for distributed power supply, TDGmaxIs the maximum number of hours of power generation, M, of the distributed power supplyDGTotal number of distributed power sources, P, for access to the distribution networkLATo total capacity of the grid, P∑DGIs the total active output of the distributed power supply.
4. The method for planning a distributed power source in a power distribution network based on maximum voltage correlation entropy of claim 1, wherein the function of the voltage quality in the step 1(c) is as follows:
wherein, UMCCIs the voltage quality, i.e. the maximum associated entropy of the node voltage, n is the number of nodes, UiSigma is a kernel function parameter, which is a per unit value of each node voltage.
5. The method for planning a distributed power supply in a power distribution network based on maximum voltage correlation entropy of claim 1, wherein the function of the power flow constraint of the power distribution network in the step 1(d) is as follows:
wherein P isiFor active injection at node i, QiFor reactive injection at node j, UiAnd UjThe voltage amplitudes at nodes i, j, G, respectivelyijConductance for branch ij, BijSusceptance, θ, of branch ijijIs the voltage phase angle difference between nodes i and j;
the function of the equipment capacity constraint in step 1(d) is:
wherein P isDGiFor active power of distributed power, QDGiFor reactive power of distributed power, PDGimaxFor rated active power, Q, of a distributed power supplyDGimaxRated reactive power for the distributed power supply;
the function of the node voltage constraint in step 1(d) is:
Uimin≤Ui≤Uimax(9)
wherein U isiIs the node voltage, UiminIs the lower limit of the node voltage, UimaxIs the upper node voltage limit.
6. The method for planning a distributed power source in a power distribution network based on the maximum voltage correlation entropy of claim 1, wherein the cross operator in the step 2.2 is as follows:
wherein P is1And P2Two father individuals randomly selected from the initial population respectively; p1 new、P2 newThe filial generation generated after the operation of the cross operator corresponds to the new individual; omega1、ω2Is [0, 1]]And (4) selecting the parameters randomly.
7. The method for planning a distributed power source in a power distribution network based on the maximum voltage correlation entropy according to claim 1, wherein the mutation operator in the step 2.3 is specifically:
wherein V is selected variant individual, V' is variant individual, sign randomly takes 0 or 1, bsupAnd binfRespectively an upper bound and a lower bound of the parameter value, r is [0, 1]]A random number generated above;
for population evolution markers, where gcIs the current evolutionary algebra of the population, gmIs the maximum evolutionary algebra of the population.
8. According toThe method for planning a distributed power source in a power distribution network based on maximum voltage correlation entropy of claim 1, wherein the selection probability P in the step 2.5 isiThe following formula:
wherein JiAnd the fitness value corresponding to the population i.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710671973.1A CN107612016B (en) | 2017-08-08 | 2017-08-08 | Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710671973.1A CN107612016B (en) | 2017-08-08 | 2017-08-08 | Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107612016A CN107612016A (en) | 2018-01-19 |
| CN107612016B true CN107612016B (en) | 2020-01-14 |
Family
ID=61064596
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710671973.1A Active CN107612016B (en) | 2017-08-08 | 2017-08-08 | Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107612016B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111854822A (en) * | 2020-06-24 | 2020-10-30 | 浙江大学 | Semiconductor process data correction method based on correlation entropy and deep neural network |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108448609B (en) * | 2018-03-01 | 2020-05-22 | 华南理工大学 | Distribution network loss allocation method considering DG equivalent capacity utilization hour coefficient |
| CN108564204B (en) * | 2018-03-23 | 2022-01-07 | 西安理工大学 | Least square support vector machine electricity quantity prediction method based on maximum correlation entropy criterion |
| CN109038653A (en) * | 2018-07-18 | 2018-12-18 | 西安理工大学 | A kind of power distribution network optimizing operation method for pursuing the consumption of distributing wind electricity generating |
| CN108964124B (en) * | 2018-07-19 | 2021-03-16 | 国网江苏省电力有限公司宿迁供电分公司 | Distributed power supply grid-connected optimization configuration method for counting and electricity price response in incremental power distribution network |
| CN109409616A (en) * | 2018-11-22 | 2019-03-01 | 南方电网科学研究院有限责任公司 | A kind of distributed generation resource addressing constant volume method and Related product |
| CN111798494B (en) * | 2020-05-22 | 2023-08-22 | 浙江工商大学 | Maneuvering target robust tracking method under generalized correlation entropy criterion |
| CN111861094A (en) * | 2020-05-28 | 2020-10-30 | 南京邮电大学 | Comprehensive evaluation method for security situation of distributed power supply access power distribution network |
| CN111896038B (en) * | 2020-06-24 | 2021-08-31 | 浙江大学 | Semiconductor process data correction method based on correlation entropy and shallow neural network |
| CN112365196A (en) * | 2020-12-03 | 2021-02-12 | 国网河北省电力有限公司信息通信分公司 | Spark-based power supply planning method for distributed improved differential evolution algorithm |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104376410A (en) * | 2014-11-06 | 2015-02-25 | 国家电网公司 | Planning method for distributed power source in power distribution network |
| CN105207195A (en) * | 2015-06-19 | 2015-12-30 | 河海大学 | Locating and sizing method for distributed power supply in power distribution network |
| CN106295885A (en) * | 2016-08-11 | 2017-01-04 | 中国电力科学研究院 | Active distribution network based on active management pattern associating planing method |
| CN106803130A (en) * | 2016-12-15 | 2017-06-06 | 国网安徽省电力公司淮北供电公司 | Distributed power source accesses the planing method of power distribution network |
| CN107330302A (en) * | 2017-07-10 | 2017-11-07 | 无锡职业技术学院 | The biological die body reconstructing method of joint Memetic algorithms and S system |
-
2017
- 2017-08-08 CN CN201710671973.1A patent/CN107612016B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104376410A (en) * | 2014-11-06 | 2015-02-25 | 国家电网公司 | Planning method for distributed power source in power distribution network |
| CN105207195A (en) * | 2015-06-19 | 2015-12-30 | 河海大学 | Locating and sizing method for distributed power supply in power distribution network |
| CN106295885A (en) * | 2016-08-11 | 2017-01-04 | 中国电力科学研究院 | Active distribution network based on active management pattern associating planing method |
| CN106803130A (en) * | 2016-12-15 | 2017-06-06 | 国网安徽省电力公司淮北供电公司 | Distributed power source accesses the planing method of power distribution network |
| CN107330302A (en) * | 2017-07-10 | 2017-11-07 | 无锡职业技术学院 | The biological die body reconstructing method of joint Memetic algorithms and S system |
Non-Patent Citations (3)
| Title |
|---|
| Locating and Capacity Determination of Distributed Generations Using None-Dominated Sorting Genetic Algorithm;Farzad Banihashemi 等;《IEEE GRENOBLE POWERTECH (POWERTECH)》;20131104;第1-8页 * |
| 基于启发式的配电网重构Memetic算法;胡亚南;《科技创新》;20161231(第29期);第8-9页 * |
| 基于混合蛙跳算法的分布式风电源规划;张沈习 等;《电力系统自动化》;20130710;第37卷(第13期);第76-82页 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111854822A (en) * | 2020-06-24 | 2020-10-30 | 浙江大学 | Semiconductor process data correction method based on correlation entropy and deep neural network |
| CN111854822B (en) * | 2020-06-24 | 2021-06-29 | 浙江大学 | Semiconductor process data correction method based on correlation entropy and deep neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107612016A (en) | 2018-01-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107612016B (en) | Planning method of distributed power supply in power distribution network based on maximum voltage correlation entropy | |
| CN109995075B (en) | Dynamic reconstruction method for active power distribution network containing distributed power supply | |
| Li et al. | Coordinated scheduling for improving uncertain wind power adsorption in electric vehicles—Wind integrated power systems by multiobjective optimization approach | |
| CN106972504B (en) | Interval reactive power optimization method based on genetic algorithm | |
| CN111639870A (en) | Power transmission network multi-target flexible planning method and system considering source load uncertainty | |
| CN110889603A (en) | Power system economic dispatching method considering wind power correlation based on PCA-Copula theory | |
| CN106803130B (en) | Planning method for distributed power supply to be connected into power distribution network | |
| CN109038545B (en) | Power distribution network reconstruction method based on differential evolution invasive weed algorithm | |
| Safari et al. | Optimal DG allocation in distribution network | |
| CN110429636B (en) | Static voltage stabilization fault screening and sorting method | |
| CN112994017A (en) | Distributed photovoltaic power supply site selection optimization method based on power distribution network probability load flow calculation | |
| Javidsharifi et al. | Probabilistic model for microgrids optimal energy management considering AC network constraints | |
| Basu et al. | Reliability study of a micro grid system with optimal sizing and placement of DER | |
| Nayak et al. | Optimal placement and sizing of distributed generation in radial distribution system using differential evolution algorithm | |
| Xiao et al. | Optimal sizing and siting of soft open point for improving the three phase unbalance of the distribution network | |
| Ebrahimi et al. | Probabilistic optimal planning in active distribution networks considering non‐linear loads based on data clustering method | |
| CN112446175B (en) | Distributed power supply site selection optimization method based on power distribution network probability load flow calculation | |
| Saric et al. | Optimal DG allocation for power loss reduction considering load and generation uncertainties | |
| Lee et al. | Advanced Probabilistic Power Flow Method Using Vine Copulas for Wind Power Capacity Expansion | |
| Karthikeyan et al. | Application of self-adaptive differentialevolution algorithm to generation expansion planning problem | |
| Hosseini et al. | Optimal placement of dgs in distribution system including different load models for loss reduction using genetic algorithm | |
| CN110571791B (en) | Optimal configuration method for power transmission network planning under new energy access | |
| Wang et al. | An alternative method for estimating wind-power capacity credit based on reliability evaluation using intelligent search | |
| Ali et al. | Distributed generation sizing and placement using computational intelligence | |
| CN111860617A (en) | Comprehensive optimization operation method for power distribution network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |